Thomas Halpin

• Managed a team of 8, providing guidance and direction to ensure successful execution of projects.
• On a joint project with the Booth School of Business and the YPO organization, I was the Wine Access tech lead to create an LLM that would write our content-based emails. The end result of this first trial was a tie between our human writers and our AI, but future versions are sure to win.
• Led the design and implementation of a behavioral model that effectively examined customer's interactions with the website and emails, leading to precise wine recommendations and accurately targeted emails.

• Designed and led the implementation of a best-in-class wine logistics program that allowed us to analyze temperature conditions along our delivery routes, ensuring that our customers' wine orders were released only when the weather was safe.
• Transitioned the website to a serverless architecture using Amazon Lambda. This project was completed weeks before the pandemic lockdown, resulting in us being able to handle a 400% increase in traffic with no increase in costs, nor any downtime.
• Leveraged my wine knowledge and led the development of an AWS Personalize-powered wine recommendation engine. The end result was a 62% higher revenue where this wine recommendation engine was deployed. parts of the website End result was 62% higher revenue where this wine recommendation engine was deployed.

• Led the technology side of the company through a lean transitional period, being the only developer/manager/qa person for the better part of the year. I kept the website running and sales flowing with a minimum of resources.
• Transitioned from standard MySQL to Amazon Aurora, netting a savings of roughly 50% on database costs (dependent on volume) over traditional MySQL AWS RDS.

• Built an automated testing tool using PIL to interface with Litmus’ API, enabling the QA team to rapidly see
  how changes to the email templates looked in dozens of browser/version combinations.
• Built multiple web scrapers to gather wine reviews from several wine websites for incorporation into our Buyer’s Guide, and used clever string matching algorithms (Levenstein and others) to match reviews from those websites with
  specific products in our database.
• Primary software engineer in requirements gathering, leveraging a decade of wine business-centric knowledge to deliver higher quality deliverables to Marketing and Product teams.

• Relationship manager and business analyst for several global banks, rolling out FATCA/KYC software.
• Responsible for extensive requirements gathering from stakeholders, and working with compliance stakeholders to develop, architect, and implement solutions.

* Responsible for management of company relationships with over 150 wine retailers.

* Responsibilities include 24/7 time-critical technical support and maintenance, training staff at respective stores about how best to utilize the company's proprietary software, and working with existing clients to update and modify current software to respond to business needs.

• Managed a team (max headcount of 5) that was charged with both identifying new wine stores to add to our network, as well as mapping wines in store’s catalogs to our database.

* Worked as a key member of the development team to thoroughly test, debug, and deploy weekly software updates.

* Developed application code in TCL, Perl, and Ruby, as well as front-end javascript development.

* Worked on natural language processing algorithms to map hundreds of stores' inventory catalogs to specific items in our database. The end result of this work resulted in the best comparative shopping engine on the web for this industry.